Mathematica Aeterna, Vol. 1, 2011, no.

05, 297 - 304

Two Degree non Homogeneous Differential Equations

with Polynomial Coefficients
Nikos Bagis
Aristotelian University of Thessaloniki
Department of Computer Sciences
Thessaloniki Greece
bagkis@hotmail.com
Abstract
We solve some forms of non homogeneous differential equations using a new function ug which is integral-closed form solution of a non
homogeneous second order ODE with linear coefficients. The non homogeneous part is an arbitrary function of L2 (R). Using this function
ug as a base we give in closed integral form solutions of several two
degree DE.

Mathematics Subject Classification: 34A05, 34A25, 34A35

Keywords: Differential equations, Linear, Non homogeneous, Closed form,
Solutions, Fourier transforms

Introduction

We will solve the equation

(a1 x + b1 )f (x) + (a2 x + b2 )f (x) + (a3 x + b3 )f (x) = g(x)

(1)

where f , g L2 (R) and a1 , a2 , a3 , b1 , b2 , b3 are constants in R.

We call the solution ug , and using this solution we try to solve other general
differential equations.
The solution of (1) (function ug ), follow using the Fourier transform and its
properties. The result is a quite complicated integral. However the Fourier
transform of this integral is not so complicated and satisfies as we will see a
simple first order differential equation. We also consider and solve completely
a class of differential equations having two degree polynomial coefficients.
We also give examples and applications. Special cases of ug are known functions, such as Hermite polynomials, Bessel functions, confluent hypergeometric
functions and other.

298

Nikos Bagis

Preliminary Notes

Let the Fourier Transform of a function f of L2 (R) is

fb() =

f (t)eit dx

the Inverse Fourier Transform is

f (x) =

1
2

Then it is known (integration by parts)

Z

f (x)A(x)eix =

(2)
(3)

f (x)A(x)eix dx =

f (x)A (x)eix dx + (i)

+2(i)

(n) )() = (i)n fb().

(fd

f (x)xn eix dx = in (fb)(n) ().

Z

fb()eix d

ix

f (x)A (x)e

f (x)A(x)eix dx.

(4)

f (x)A (x)eix dx+

dx + (i)

f (x)A(x)eix dx.

(5)

(see [2]).

Main Results

These are the main results of the paper.

Theorem 3.1 When f , g L2 (R) and lim|x| |f (x)x2+ | = 0, > 0,
equation (1) can reduced to
(ia1 2 + a2 + ia3 )
which is solvable.

fb()
+ (b1 2 2ia1 + ib2 + a2 + b3 )fb() = gb() (6)
d

Proof.
Take the Fourier Transform in both sides of (1) and use relations (2),(3),(4),(5).
We immediately arrive to (6). But (6) is completely solvable DE of first degree.

299

Two Degree non Homogeneous Differential Equations

After solving it we use the inverse Fourier transform to get the solution, which
we call it
f (x) = ug [{a1 , a2 , a3 } , {b1 , b2 , b3 } ; x] .
(7)
and satisfies (6).
Note. When g(x) = 0, the solution is very simple as someone can see is
uc0 [{a1 , a2 , a3 } , {b1 , b2 , b3 } ; ] =
= exp

b1 2 2ia1 + ib2 + a2 + b3
d
ia1 2 + a2 + ia3

(8)

Theorem 3.2 The DE

(ax2 + bx + c)y (x) + (kx2 + lx + m)y (x) + (tx2 + rx + s)y(x) = g(x) (9)
is solvable if
a (4a bl l + 2am)
b2 + 2ac b

(10)

2b2 2b
bm m br r
b2 t
ct bt
+
l+

+
+
2+ 2
c
c
c
c
2a
2a
2a
a
2a

(11)

k=
s = 6a

and = b2 4ac
The solution is

y(x) =

a1 x + b1
ug [{a1 , a2 , a3 } , {b1 , b2 , b3 } ; x]
ax2 + bx + c

(12)

where
a1 =

ab3 (b + )
(b + ) k + 2a(2a l + s)

(13)

a2 =

b3 (b + ) k
(b + ) k + 2a(2a l + s)

(14)

a3 =

b3 (b + ) t
(b + ) k + 2a(2a l + s)

(15)

b1 =

2ab3 c
(b + ) k + 2a(2a l + s)

(16)

b2 =

b3 (b + ) (4a2 + bk + k 2al)
2a (4a2 + bk + k 2al + 2as)

(17)

300

Nikos Bagis

Proof. We proceed as in Proposition 1.

Set
y(x) =

a1 x + b1
Y (x)
ax2 + bx + c

(18)

We arrive in a equation
(a1 x + b1 )Y (x) + (a2 x + b2 )Y (x) + (a3 x + b3 )Y (x) = g(x),

(19)

in which the relation of the coefficients of the two equations (9) and (19) is:
b=

ab21 + a21 c
aa2
ab2 a2 c
aa3
,k=
, l = 2a +
+
,t=
a1 b1
a1
a1
b1
a1

2ab21 + a1 b2 c
aa2 b1 + ab1 b3 + a1 a3 c
ab1 b2 + a1 b3 c
,r=
,s=
a1 b1
a1 b1
a1 b1
Solving the above system we get the proof.
i) Theorem 2 prove that, a two degree ODE with second degree polynomial
coefficients is 6-th parameter solvable DE and admits a general solution y(x) =
ug (x) in a closed integral form. Thus if we give any values we want in (9) then
it is solvable always except for two values of the coefficients (here) k and s,
which are determined by the others.
ii) A way to find some special cases of solutions of the DE (9) is with the
command SolveAlways of the program Mathematica.
Example 1. The equation

!


207

25
69
2

2
x + 3x + 1/3 y (x) + x
+ 2x + 3 y (x)+
2
!
233 + 41 69
x2
=0
(20)
+x+
+y(x)
2
12
m=

is equivalent to the second order, first degree polynomial coefficients ODE

(95 23 69 + (9 + 69)x)Y (x) + 2(2 69 + (69 9 69)x)Y (x)+

+(2 + (9 + 69)x)Y (x) = 0,

(21)
which is solvable according to Theorem 1.

9(2 + (9 + 69)x)

y(x) =
(95 + 23 69)(1 + 9x + 3x2 )
hn

o n
oi
u0 6 + 69, 2(69 9 69), 9 69 , 2, 4 69, 95 + 23 69

where u0 is given by (8).

301

Two Degree non Homogeneous Differential Equations

Theorem 3.3 The DE

"

"

h
i
2ab1
ab1
y (x)+ tx2 + rx + s y(x) = g(x),
x y (x)+ kx2 + (2a + s)x +
ax +
a1
a1
(22)
with g L2 (R), have a closed integral form solution
2

y(x) = ug

"(

) (

a1 k a1 t
a1 s a1 k + a1 r
a1 ,
, b1 ,
;x
,
,
a a
a
a

(23)

Applications in the case of linear coefficients

1.
The homogeneous of the differential equation
2xy (x) + y (x) 2y(x) = g(x)

(24)

2xy (x) + y (x) 2y(x) = 0

(25)

is
and have solutions

y1 (x) = cosh(2 x) , y2 (x) = sinh(2 x)

With our method (24) have solution
y(x) = ug [{2, 0, 0} , {1, 1, 2} ; x]
where

Z w i/x
ei/w
ie G(x)
b

y (w) = 3/2 C1 +
dx
w
2 x
c

and G(x) = gb(x).

2.
The homogeneous of the differential equation

y (x) 2xy (x) + 2ny(x) = g(x)

(26)

y (x) 2xy (x) + 2ny(x) = 0

(27)

is
and have solution


n 1
y1 (x) = Hn (x), y2 (x) = 1 F1 , ; x2
2 2

302

Nikos Bagis

With our method (26) have solution

y(x) = ug [{0, 2, 0} , {1, 0, 2n} ; x]
where

yb(w) = ei(1+n) arctan( w ) w 1+n 4 + w 2

2

 1 n
2

iei(1+n) arctan( x ) G(x)xn (2i + x) 4 + x2

2

and G(x) = gb(x).

3.
The homogeneous of the differential equation

[C1 +

 1 (1+n)
2

dx]

xy (x) + y(x) = g(x)

(28)

xy (x) + y(x) = 0

(29)

is
and have solution
y(x) =

i
x C1 J1 (2 x) + C2 Y1 (2 x)

With our method (28) have solution

y(x) = ug [{1, 0, 0} , {0, 0, 1} ; x]
where
yb(w)

1 i
= 2 e w C1 +
w

xi

ie

G(x) dx

and G(x) = gb(x).

Note.
In the above examples if we extract the general solutions and then set in the
equations G(x) = 0, we arrive to the reduced homogeneous solution.

Applications for the two degree polynomials coefficients

1.
The homogeneous of the differential equation
x2 y (x) + [kx2 + x(2 + s)]y (x) + (tx2 + rx + s)y(x) = g(x)

(30)

x2 y (x) + [kx2 + x(2 + s)]y (x) + (tx2 + rx + s)y(x) = 0

(31)

is

303

Two Degree non Homogeneous Differential Equations

and have solution

y(x) =
C1 kx/2k2 4tx/2
e
U
=
x

2r + k(2 + s) + s k 2 4t

, s, x k 2 4t +
2 k 2 4t

2r k(2 + s) s k 2 4t
C2 kx/2k2 4tx/2
2

L
+ e
, 1 + s, x k 4t
x
2 k 2 4t

Where Lan (x) = L(n, a, x) is the generalized Laguerre polynomial and U(a, b, x)
is confluent hypergeometric function.
With our method (30) have solution
y(x) =

a1
ug [{a1 , a1 k, a1 t} , {0, a1 s, a1 k + a1 r} ; x]
x

2.
The equation
x(1x)y (x)+(kx2 +lx2kl)y (x)+[tx2 +(s+2+2k+lt)x+s]y(x) = g(x)
have solution
y(x) =

a1

1x

ug [{a1 , a1 k, a1 t} , {0, 2a1 a1 k a1 l, 2a1 a1 k a1 l + a1 s} ; x]

Example of the above equation we get if we choose b1 = 0, a2 = 0, b2 =
2a1 a1 l, a3 = a1 t, b3 = 2a1 a1 l + a1 s and k = 0, then
i(2+ls)

or

yb(w) = e

y(x) =
is the solution of

w 4l C1 +

ie

i(2+ls)
x

G(x)x2+l
dx
a1

a1
ug [{a1 , a2 , a3 } , {b1 , b2 , b3 } ; x]
1x

(1 x)xy (x) + (lx 2 l)y (x) + [tx2 + (2 + l t)x + s]y(x) = g(x)

3.
The equation
x(1 x)y (x) + (kx2 + lx + m)y (x) + (tx2 + rx + s)y(x) = g(x)
is solvable with the ug functions when k + l + m + 2 = 0.

(32)

304

Nikos Bagis

References
[1] N.N. Lebedev, Special Functions and their Applications, Dover Publications, Inc. New York (1972)
[2] A. Papoulis, The Fourier Integral and its Applications, McGraw-Hill Publications, New York (1974)
[3] Murray R. Spiegel, Schaums Outline of Theory and Problems of Fourier
Analysis with Applications to Boundary Value Problems, McGraw-Hill,
Inc, (1974).
[4] Greek Graduate books in Differential Equations
Received: July, 2011